Google Cloud Text-to-Speech Reviews

We use Speech-to-Text and Text-to-Speech to be able to talk to our users. We have an AI meaning engine that back-ends that. Once we get the speech, we can tell what it means. That's our use case.

When we tested Speech-to-Text a few years ago, it was better than the equivalent IBM products because it was more accurate. But it's not by any means 100% accurate, and we have to correct for those errors in our AI software.

It's using neural networks and that stochastic processing is 70-75% accurate. It gets it wrong too often, and we really don't like that. Since I personally work with this, I don't like that a lot. But they seem to be the best game in town right now.

There's nobody else out there doing STT and TTS that is as good as them. There are several competitors trying, including Nuance and IBM, but their solutions are not very good, at least not as good as Google Cloud Text-to-Speech.

Google Cloud Text-to-Speech is 100 out of 100 when it works, and when it doesn't work, which is fairly often, it gets a zero. It doesn't fail gracefully, it doesn't fail in an expected way, it just fails. It just stops working and then we get errors, and then my development people have to try and figure out what those are and do something about it.

If you want to talk to your computer, then in our view, using Google Cloud Text-to-Speech is the way to do it. But it's not without problems. The difference between how it works in a Windows environment on a development machine and how it works up in the cloud are unsettling and costly because we're spending expensive engineering time trying to figure it out.

The support is inadequate. We are dealing with them on our development talk today. There's a lot of finger-pointing going on in terms of whose problem it is. Moving our stuff up to the Google Cloud and getting it to work just as well as it does on people's development machines is problematic. Their support for that, even though we paid for it, isn't really very helpful. That's prevalent in the computer business. You need to have your own experts, otherwise you're really in trouble.

The product is an eight out of 10. The support is at best a five. We have to write certain features ourselves because their offerings aren't very powerful.

When I don't have a problem, it works pretty well, better than anybody else. But when I do have a problem, I'm severely impacted. It takes a lot of time and money to go back and fix it.

What has gotten better with Google Cloud Text-to-Speech is their stuff sounds so natural, it really brings a smile to my face. I wish their support would be better.

I have been using this solution my whole life.

Moving our stuff up to the Google Cloud and getting it to work just as well as it does on people's development machines is problematic. The support for that, even though we paid for it, isn't really very helpful.

On development machines and laptops, it works great. It works out of the box. That's terrific and one of the reasons we went with them. Up in the cloud though, we encounter error after error.

The product is the best of breed as far as we're concerned right now. When I don't have a problem, it works pretty well, better than anybody else. But when I do have a problem, I'm severely impacted. It takes a lot of time and money to go back and fix it.

Moving our stuff up to the Google Cloud and getting it to work just as well as it does on people's development machines is problematic. The support for that, even though we paid for it, isn't really very helpful. Up in the cloud, we encounter error after error.

We just got error messages and we couldn't get any useful responses from customer support. Their support is inadequate. The support is bad.

Neutral

We evaluated IBM and Nuance. I personally use AWS, but we don't use their cloud. We use the Google Cloud. We're not buying it from AWS, we're buying it directly from Google.

The accuracy is the most important factor. It doesn't matter what you do with the utterance if you don't recognize it. These other products didn't recognize it as well as Google Cloud Text-to-Speech does.

We haven't done any of that yet. We're following the research at MIT of Rosalind Picard about those kinds of issues. We plan to do something with it, we just haven't quite got there yet.

Our expectation was we didn't have any other choice. We can't really say that it's well-priced or badly priced. We just didn't have another choice as far as we were concerned. The product is the best of breed as far as we're concerned right now.

Our experience is we didn't have any other choice. We can't really say that it's well-priced or badly priced. We just didn't have another choice as far as we were concerned.

IBM and Nuance were the products we evaluated.

I'm not familiar with Speaker Diarization. Regarding the custom voice creation feature, we just use their voices, which are fine for our concerns.

We haven't used it with Google's ML. However it's coming out of the box to us, we've got enough problems understanding the meaning of the utterance. We don't want to spend money on that if Google can do it.

Part of our metrics involves call abandonment and some internal metrics we've developed about understanding what to do in a conversation with an utterance and how that plays with the users. Currently, it's disappointingly bad with complex conversations. Simple queries are easy, but real human conversations need a lot of work with AI support.

Their pricing is competitive, but what matters most is that it works. The other competitors don't work well enough for us to consider them. That's just the cost of doing business.

Overall rating: 8/10

We use Google Cloud Text-to-Speech when we have an IVR and need to say something written on the API through the IVR. For example, if I'm on the IVR and need to say something to the customer that is written on the API, I need to vocalize this text. I use Google Cloud Text-to-Speech to take this text and vocalize it to the client.

Precision is the most valuable feature of Google Cloud Text-to-Speech because the text is perfectly voiced.

Google Cloud Text-to-Speech has just one female voice and one male voice in Brazil, while it has a lot of voices in other countries.

I have been using Google Cloud Text-to-Speech for two years.

Google Cloud Text-to-Speech has very good stability.

Google Cloud Text-to-Speech can be used extensively because you use and pay for it. The solution's scalability is very good.

The solution's initial setup is very simple. You have to register on Google Cloud and start using it.

I rate Google Cloud Text-to-Speech three out of ten for pricing.

The voice on Microsoft Text-to-Speech is a little bit better than the voice in Google Cloud Text-to-Speech.

Overall, I rate Google Cloud Text-to-Speech ten out of ten.

The solution is used for developing translators for chatbots.

In the past year, we won the hackathon with the chatbot and the chatbot is in two languages. We needed some solution for that translation. We didn't want to have two chatbots, one in English and one in Spanish. We wanted to have one chatbot in a multi-language format, and that is really a problem. 

I like the languages and I made a multi-translator. You can write some in Spanish and then you can see the same sentence in English, Russian, and Hindi. For learning purposes, I'm interested in romance languages. I can compare the differences between Spanish, Portuguese, Romanian, and other romance languages. 

I like how I can direct language, how I can direct the origin language, for example.

It's not complex to set up.

When you make a chatbot, you need to do it for only one language, typically. We now work in multilanguage companies, however, and we need a chatbot with smarter skills. We need a chat with the capacity of comprehension for English, Spanish, Hindi, and Chinese maybe, since we have people working in all that countries.

We had some problems with Dialogflow. 

I don't like the sentiment analysis. I don't feel like it's that realistic. 

I've been using the solution for two to three years. 

It's very stable. The translation capabilities are better than, for example, Microsoft.

We did deal with technical support in the past. We wanted to make a multilanguage chatbot, however, we can't do that. You need to set it in one language.

I'm also familiar with Microsoft, however, Google can translate language more effectively.

The solution is easy to set up. It's not difficult. I'd rate it a five out of five in terms of the ease of setup. It's very straightforward. 

The deployment is quick. It only takes a day or so. 

I don't have a realistic outlook on the cost. We are working on the migration from server to data center in Jira, and that is a lot of work.

There is a lot of money to be made in languages. If you manage to create something that's very good with translation, you're going to become very wealthy.

I'd rate the solution a seven out of ten.